2012-02-24 12:36:49Feedbacks and Runaway Warming
Chris Colose


Just an addition that was motivated by the "runaway feedback" section in the forthcoming Monckton article.  I could work on that section next if you want, or this (and the other SkS post) could simply serve as a reference.  Hope you like...


2012-02-24 13:49:16
Tom Curtis


"I wanted to revisit the subject in order to put a different spin on take a different approach to the subject of positive feedbacks."

"Spin" would be spun no end by deniers.  You may even find yourself features in a "They told us they're going to lie" section fo  Monckton presentation.

2012-02-24 14:08:05
Tom Curtis


Chris, if I understand this correctly, you are discussing the Kombayashi-Ingersol limit.  IMO, it would be worthwhile mentioning after you have fully explained the concept.  That way the large hyphenated name doesn't get in the way of following a nice clear exposition, but people can then relate that clear exposition to other discussions they may have seen on the net.


Also, did you include any other feedbacks (notably the lapse rate feedback) in your model?


Finally, and out of curiosity, I assume the effect of including other positive feedbacks would be to shift the blue curve down and to the right.  Presumably that would make it more probable that we exceed the limit that initiates a runaway greenhouse effect.  Do you know how much the Earth's albedo would need to decrease for that to be possible?  Is it possible for the melting of the Earth's ice caps to decrease albedo by that amount?


I do not think discussion of the points in the last two paragraphs belong in the post.  I am just curious because I believe Hansen continues to think a runaway greenhouse is still a potential risk, and I am wondering why.  Regardless of that, with the two modifications (previous post and first paragraph of this post) that I have suggested, I am definitely giving this post a green thumb.

2012-02-24 16:21:42
Dana Nuccitelli

Really nice Chris.  My brain is fried right now, but I was able to follow up until the very end, which is a very good sign.

My main question is when that runaway greenhouse effect might be triggered.  Since a runaway greenhouse is possible, I'm trying to relate that to Monckton's argument that high sensitivity implies runaway-style instability.  Can you add a bit about that?

For the Monckton Part 3 post, I think the best solution would be a brief summary of your points here, and then a link to the full post.  If you can come up with a brief summary I could put in the post, that would be great.  Or I can take a stab at it when my brain re-solidifies tomorrow.

2012-02-24 17:07:08
Chris Colose




The atmosphere treated is on the moist adiabat (even if just looking at a 'global average' quantity) so the lapse rate adjustment would be accounted for automatically.  There wouldn't be cloud feedbacks.

I'm on a different computer right now so I'll have to check, but I'm quite sure I did an "all-troposphere" computation, but you can plug in a constant value for an isothermal stratosphere, though it wouldn't change much here. The radiation code download is a whole lot of fortran and I'm still not quite educated in its nuances, but it's one of the most elaborate high-speed codes for this sort of purpose that is available.  It's not quite on the level of modern line by line radiation codes, and it's not a GCM (I have just a simplifed version providing access to just the radiation model), but it's been employed in the literature quite frequently.  Ray Pierrehumbert's been doing stuff along these lines in some of his recent publications, because of how he meshed this altogether for usage with his book.

As for what this radiation limit is called, there's actually a lot of subtlety to this, as it depends on the precise physical processes that occur.  I've called this the Komabayahsi-Ingersoll limit before but Nakajima back in the early 90s identified a different limit (and calledi t something else), and distinguished between troposphere and stratosphere limits, but the end result (and what I put in this post) is the same regardless. If you're interested in the technicalities, you could see the work just done this year on this. 


I'm not so sure it's pratical whatsoever, because all that would really matter is the lower of these limits, which has been identified as being somewhere around 310 W/m2 for a couple decades (e.g., in James Kastings' pioneering work in 1988, but there were papers on the runaway prior to this).  That number could increase quite a bit though for an atmosphere well below 100% relative humidity.  It is a lot tougher to get a runaway on theoretical planets with a lot of land and only a bit of ocean.  Moreover, we're several decades beyond some of the pioneering calculations, and they haven't been satisfactorily revisited in the literature. 

If that number is approximately true, then the Earth's albedo would need to be reduced from ~30% to 10% to start being worried about a runaway.  I don't really see a scenario in which that is plausible because in the high water vapor limit you have clouds and a lot of Rayleigh scattering from the atmosphere.  Most people in this field think clouds would make a runaway greenhouse effect harder rather than easier, because their infrared effect is swamped by the vast amount of water vapor in the air, so the albedo would probably win out but to date no credible calculation exists of how clouds modify the runaway picture.

  But in this limit, the surface albedo becomes almost completely irrelevant for the total planetary albedo.  Ice on Earth actually matters far less for the planetary albedo even today than people seem to think, and is more of a locally important feedback.

As for Hansen, as brilliant as he is, I don't think he understands the runaway greenhouse effect as planetary scientists do.  I spent my summer at GISS and I never had the opportunity to talk to him (seen him only once on an elevator).  Ray Pierrehumbert has blasted him a few times at RC (and I've talked to him privately about it), and in my own reading of the matter, Hansen's claims are completely indefensible.  CO2 doesn't change the radiation limit all that much even if it changes the surface temperature (see Fig. 5 in the arXiv link above).  There's been some older calculations (e.g., Kasting and Ackerman, 1986) that have dumped pretty much all the CO2 in the Earth's rocks into the atmosphere and you get a super-hot atmosphere, well above the modern boiling point, but you don't actually lose the oceans.  You can get hotter than 373 K and still have liquid water because the atmospheric pressure also increases (due to water vapor going up), so for Earth you actually need to look more toward the critical point of water on a thermodynamic phase diagram (647 K, and about 220 bars of pressure).  Adding all the CO2 you want doesn't get you there, especially because a super-dense CO2 atmosphere also starts to increase the albedo a lot eventually by solar scattering.

2012-02-24 17:32:41
Glenn Tamblyn


Chris. Interesting post. Let me play devils advocate and pre-empt comments that you might get from this. The most noticable change in going from No Amosphere to CO2 is the curve shifts down but is still an increasing slope. Adding Water however produced a curve with reducing slope. So what happens to the CO2 & Water curve if we just keep adding CO2? Will that shift that curve down but retain its current slope, giving seemingly a greater temperature response.

I think the answer is the logarithmic nature of CO2 forcing is what will prevent it from being serious, but this isn't clear from this post and people could possibly get the wrong take on it.

I think what I am saying is that truely massive CO2 levels might shift the curve far enough down to trigger a Venus effect but the levels involved are monstrous. It might be worth finding a way to show this.

2012-02-24 17:53:10
Tom Curtis


Chris, thanks for the explanation.  I'm not sure how reassuring it is to know that we won't have runaway greenhouse, but its theoretically possible to lift the GMST to 100 degrees C, but it is certainly good to get the details straightened out in my head.

2012-02-24 18:14:20
Chris Colose


Thanks Glenn,

In a post I did here a while ago on the subject (the Skeptic agument, "Venus doesn't have a runaway...") I took this image from a 2002 Pierrehumbert paper


That shows essentially the same thing as these plots but has some emphasis on your question.  There are two lines in that diagram that get at your question, both have 70% relative humidity, but one has 0.2 bars of CO2 compared with 100 ppm in the other case.  As I mentioned above to Tom, this additional CO2 will reduce the OLR for any given temperature (or equivalently, one will have to intersect the equilibrium point at a higher temperature).  But it doesn't change the radiation threshold at which the OLR becomes pretty flat very much (and actually, the differences there are larger than I've seen in other works).  Perhaps a more clear diagram would be Figure 5 in the arXiv link that I gave above to Tom.  That shows a few different curves with different CO2 concentrations. There's some exotic effects shown there that can occur when you have background greenhouse gases (not water vapor) and different relative humidities; for example in the Figure 5 you can get an overshoot of the radiation limit, and in certain circumstances CO2 can push you "over the edge." 

I'd be happy to get into this in the comments, and it won't be hard for me to make another plot with a different CO2 concentration than 400 ppm, but I don't want to clutter up the main post with stuff not really related to my main point.  I also added a final paragraph at the end discussing why we really don't need to worry about a runaway in the forseeable future.

2012-02-24 18:49:49
Glenn Tamblyn


I suppose that is the point I am making. I think the observation I made will be made by others. So the key question is do you want to nail this in the post, so there is no confusion, vs dealing with it in the comments. Not everyone works through the comments. And I can see the point of not wanting to add to the length of the post.

That makes it a call about what you want the post to achieve.

I hadn't come across the KI Limit stuff before, interesting. And relevent to our descendents 500 million years from now. But I don't think we have to take responsibility for that. We have enough on our plate.